Page MenuHomePhabricator

willschm (Will Schmidt)
User

Projects

User does not belong to any projects.

User Details

User Since
Aug 20 2014, 8:37 AM (261 w, 16 h)

Recent Activity

Jan 14 2015

willschm committed rL226045.
Jan 14 2015, 2:09 PM
willschm committed rL226032: Add a little_endian reference output entry for the 2007-01-07-lvsl-lvsr….
Add a little_endian reference output entry for the 2007-01-07-lvsl-lvsr…
Jan 14 2015, 12:11 PM
willschm closed D6971: add little_endian reference output for 2007-01-07-lvsl-lvsr-Regression test by committing rL226032: Add a little_endian reference output entry for the 2007-01-07-lvsl-lvsr….
Jan 14 2015, 12:11 PM
willschm retitled D6971: add little_endian reference output for 2007-01-07-lvsl-lvsr-Regression test from to add little_endian reference output for 2007-01-07-lvsl-lvsr-Regression test.
Jan 14 2015, 11:11 AM

Jan 13 2015

willschm committed rL225810: Update multiline.ll testcase to handle (ppc64le) .localentry directive.
Update multiline.ll testcase to handle (ppc64le) .localentry directive
Jan 13 2015, 10:18 AM
willschm closed D6935: Update multiline.ll testcase to handle (ppc64le) .localentry directive by committing rL225810: Update multiline.ll testcase to handle (ppc64le) .localentry directive.
Jan 13 2015, 10:18 AM
willschm updated the diff for D6935: Update multiline.ll testcase to handle (ppc64le) .localentry directive.

Using a space "{{ }}" as suggested does appear to work. (Local tests against powerpc64* and x86 both pass with this change).

Jan 13 2015, 10:03 AM
willschm committed rL225791: Add buildbot_selector entry for ppc64LE sanitizer bot.
Add buildbot_selector entry for ppc64LE sanitizer bot
Jan 13 2015, 6:44 AM
willschm closed D6934: Add buildbot_selector entry for ppc64LE sanitizer bot by committing rL225791: Add buildbot_selector entry for ppc64LE sanitizer bot.
Jan 13 2015, 6:44 AM · Restricted Project

Jan 12 2015

willschm added a comment to D6935: Update multiline.ll testcase to handle (ppc64le) .localentry directive.

Actually, wait - the point of that NOT .loc is to check for any .loc
directive, which has arguments and stuff, so .loc{{$}} is too restrictive
and wouldn't catch other .locs

maybe CHECK-NOT: .loc{{ }}

to force the space? Does that work?

Jan 12 2015, 3:11 PM
willschm retitled D6935: Update multiline.ll testcase to handle (ppc64le) .localentry directive from to Update multiline.ll testcase to handle (ppc64le) .localentry directive.
Jan 12 2015, 2:23 PM
willschm updated the diff for D6934: Add buildbot_selector entry for ppc64LE sanitizer bot.

Per comment from Alexey, renamed to sanitizer-ppc64le-linux for consistency.
Also tickles builders.py and slaves.py, as seen below.
Thanks,

Jan 12 2015, 2:14 PM · Restricted Project
willschm updated subscribers of D6934: Add buildbot_selector entry for ppc64LE sanitizer bot.
Jan 12 2015, 12:32 PM · Restricted Project
willschm retitled D6934: Add buildbot_selector entry for ppc64LE sanitizer bot from to Add buildbot_selector entry for ppc64LE sanitizer bot.
Jan 12 2015, 12:32 PM · Restricted Project

Jan 9 2015

willschm committed rL225543: Add buildbot entries for ppc64le.
Add buildbot entries for ppc64le
Jan 9 2015, 12:44 PM

Dec 18 2014

willschm accepted D6717: [Sanitizer] Fix GetRSS on Linux with non-4k pages.

This looks like a good change to me.
Thanks,
-Will

Dec 18 2014, 8:10 AM

Dec 17 2014

willschm closed D6696: enable P8Model entry.

committed in revision 224452.

Dec 17 2014, 12:03 PM
willschm committed rL224452: Enable the P8Model entry.
Enable the P8Model entry
Dec 17 2014, 11:57 AM

Dec 16 2014

willschm updated subscribers of D6696: enable P8Model entry.
Dec 16 2014, 2:59 PM
willschm retitled D6696: enable P8Model entry from to enable P8Model entry.
Dec 16 2014, 2:57 PM

Dec 3 2014

willschm closed D6358: [PowerPC] Add Power8 Instruction Scheduling Tablegen.

Closed by commit rL223257 (authored by @willschm).

Dec 3 2014, 10:47 AM

Dec 2 2014

willschm updated the diff for D6358: [PowerPC] Add Power8 Instruction Scheduling Tablegen.

Thanks for the reviews.
Cosmetic updates as recommended by Hal.
Removed the erroneous trailing ",0>" bits. That was simply sloppy copy/paste adjustments on my part.

Dec 2 2014, 8:32 AM

Nov 26 2014

willschm updated the diff for D6358: [PowerPC] Add Power8 Instruction Scheduling Tablegen.

Updated with P8_LU* units added, units renamed to match documentation, and comments added/sprinkled throughout.
A notable thanks to Pat Haugen who provided guidance on a few of these changes.

Nov 26 2014, 1:58 PM
willschm added inline comments to D6410: [PowerPC] Fix unwind info with dynamic stack realignment.
Nov 26 2014, 7:59 AM
willschm added a comment to D6410: [PowerPC] Fix unwind info with dynamic stack realignment.

This looks straightforward.. A cosmetic comment inline, but I've no issues with the patch.
I defer to Uli/Hal for their approval.

Nov 26 2014, 7:11 AM

Nov 21 2014

willschm added a comment to D6358: [PowerPC] Add Power8 Instruction Scheduling Tablegen.

I've been discussing a few related details with Pat Haugen on the side. I'll make a few updates and repost early in the coming week.

Nov 21 2014, 2:13 PM
willschm retitled D6358: [PowerPC] Add Power8 Instruction Scheduling Tablegen from to [PowerPC] Add Power8 Instruction Scheduling Tablegen.
Nov 21 2014, 12:04 PM

Nov 6 2014

willschm closed D6044: [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64*.

Closed by commit rL221457 (authored by @willschm).

Nov 6 2014, 7:09 AM

Nov 5 2014

willschm updated the diff for D6044: [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64*.

Add a testcase that covers validation of address mapping for both old and new kernels.
Also some cosmetic touch-ups per previous feedback.

Nov 5 2014, 2:56 PM
willschm closed D6043: [compiler-rt] add ppc64le target arch.

Closed by commit rL221356 (authored by @willschm).

Nov 5 2014, 8:46 AM

Nov 3 2014

willschm added a comment to D6082: [Sanitizer] Update GetMaxVirtualAddress for PowerPC64 Linux.
In D6082#7, @kcc wrote:

I am always reluctant to support any old OS (and > 2+ years is old) unless there is a specific request and someone is willing to do the support work.
My team currently does't test on PowerPC at all. So, the simplest solution (i.e. only support 46-bit and assert at run-time if the OS is old) seems best for me.

Nov 3 2014, 3:27 PM
willschm added a comment to D6044: [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64*.
In D6044#11, @foad wrote:

I can confirm that this works for me on big-endian PPC64 Fedora 19.

But the comments here: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55975#c27 suggest that with "ulimit -s unlimited", the stack might not be right up near the 1<<44 or 1<<46 byte limit, so maybe it would be a bit safer to do something like:

if (GET_CURRENT_FRAME() < 1ULL << 44)
  return (1ULL << 44) - 1;
else
  return (1ULL << 46) - 1;

In practice, on my machine, this doesn't seem to be a problem. Even with "ulimit -s unlimited" the stack is still very close to 1<<46.

Nov 3 2014, 3:21 PM
willschm added a comment to D6082: [Sanitizer] Update GetMaxVirtualAddress for PowerPC64 Linux.

Hi Jay,

Can you see if the updated patch from ( http://reviews.llvm.org/D6044 ) works for you.   It should work for both old and new kernels, etc.
Nov 3 2014, 2:12 PM
willschm added a comment to D6044: [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64*.
In D6044#5, @kcc wrote:

I wonder if you can add a powerpc-specific test for this.

Nov 3 2014, 1:55 PM
willschm updated the diff for D6044: [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64*.

Reworked to use GET_CURRENT_FRAME and MostSignificantSetBitIndex helper functions per previous comments.

Nov 3 2014, 1:41 PM

Oct 30 2014

willschm retitled D6044: [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64* from to [compiler-rt] fix and simplify GetMaxVirtualAddress for ppc64*.
Oct 30 2014, 3:37 PM
willschm retitled D6043: [compiler-rt] add ppc64le target arch from to [compiler-rt] add ppc64le target arch.
Oct 30 2014, 3:25 PM

Aug 20 2014

willschm added a comment to D4977: critical-anti-dependency breaker: don't use reg def info from kill insts (PR20308).

Changes look OK to me. I would defer final say/ technical commentary to Hal, (versus myself). Hal should have a better understanding of all the parts here than I do.

Aug 20 2014, 9:10 AM
willschm updated subscribers of D4977: critical-anti-dependency breaker: don't use reg def info from kill insts (PR20308).
Aug 20 2014, 8:38 AM